Variability of building energy performance at a scale: conformity of predictive and synthesis with explanatory modelling practices
Doctoral thesis
Permanent lenke
https://hdl.handle.net/11250/3032487Utgivelsesdato
2022Metadata
Vis full innførselSamlinger
Sammendrag
Throughout its existence, humankind invented countless means and practices to design, construct and equip buildings. The number of ways to use these buildings, with or without strictly following their initially intended purpose, is even more significant. The historical and the anticipated future evolution of buildings at a varying pace further amplifies their diversity. An already complex phenomenon of building energy use is hence further entangled. Substantial variations also stem from you, my reader, whose lifestyle and occupancy patterns often cause a logical nightmare for the energy analysts like myself. These are some of the challenges in large-scale building energy research. This discipline intends to mediate the transition to a more sustainable built environment with the associated energy supply systems. Since the discipline's inception, the inherent modelling practices follow either a bottom-up or top-down approach. These two seemingly incompatible paradigms not only address the subject matter in a radically distinct manner but differ substantially in their accuracy, sensitivity, transferability, versatility, computability and usability. So far, both approaches have been used independently, concerned with leveraging their advantages and, generally, overlooking the limitations of one or another. It was, however, expected that the best interests of practice and policymaking necessitate a synergy or a combination of approaches rather than their application individually.
Seeking ways to complement bottom-up and top-down approaches laid the foundation for the thesis you are holding. The analysis of modelling purposes, targeted system's complexities, model's characteristics and the associated uncertainties were expected to provide meaningful answers. It was also understood that, under the discipline's quest for accurate prediction, explanatory modelling had been largely overlooked. Formulating and testing the causal theories can improve the understanding of building energy performance, the means to mediate it and aid with developing better predictive models. Therefore, examining the interplay of explanatory and predictive modelling, bottom-up and top-down, is another objective of this thesis. It does so through: i) four research papers that attempt to answer why and to which extent the phenomenon varies beyond its best estimates; ii) a case study that exemplifies and examines the conformity of the modelling results obtained with bottom-up and top-down reasoning. These involve a collection of instruments from detailed building energy performance simulation, known as white-box methods, and their cousins of somewhat darker shades. The latter, in this work, consists of the methods of probabilistic programming that involve statistical hypothesis testing, univariate density estimation and Monte-Carlo simulation. The methods of combinatorial analysis and numerical optimisation are applied when necessary. The modelling principles consider numerous building types, design characteristics, energy supply solutions, occupancy-related tendencies and geopolitical contexts. The findings are based on and supported by experimental data, which is, together with the essential analytical instruments, made available in Built Stock Explorer (https://buildingstockexplorer.indecol.no/). To a large extent, this research software enables reproducing/replicating our results, should you be curious about doing that. The Explorer is written in Python, a lingua franca of today's scientific computing, and evolves to facilitate an interactive built stock energy analysis and the relevant (statistical) modelling.
It is shown in this study that numerically similar built stock energy model results are achievable with either bottom-up or top-down model design. Mutual verification of the model performance in such a way can elevate the confidence of the decision making based on them. Furthermore, to prevent misleading urban developments suggested by poorly performing models. Given the importance of mediating building energy efficiency at all levels of governance, mutual verification of built stock energy modes is advocated as the means for more effective and timely achievement of energy and environmental targets. The complexities, diversity, scale and dynamics associated with building energy use at the built stock level motivate model parsimony. Explanatory modelling may inform more rational and better performing predictive model design. Also, explanatory modelling practices are expected to find applications in discovering new causal relationships, empirically validating the existing knowledge and monitoring the evolution of building energy performance at a scale. Better awareness about the phenomenon and better performance of the models at predicting it further elevate the demand for empirical data quantity and quality. This thesis advocates robust study design, data accessibility, and transparency to address the latter.
Therefore, this academic work contributes to the body of knowledge available in large-scale building energy research by focusing on and articulating the need for synthesis between various modelling practices instead of further diverging them. To the best interest of the discipline and the objectives it pursues.
Består av
Paper 1: Zhuravchak, Ruslan; Nord, Natasa; Brattebø, Helge. Control strategy for battery-supported photovoltaic systems aimed at peak load reduction. E3S Web of Conferences 2019 ;Volum 111. https://doi.org/10.1051/e3sconf/201911105027 This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0 (CC BY 4.0)Paper 2: Zhuravchak, Ruslan; Nord, Natasa; Brattebø, Helge. Influence of emerging technologies deployment in residential built stock on electric energy cost and grid load. IOP Conference Series: Earth and Environmental Science (EES) 2019 ;Volum 352.(1) s. 1-10 https://doi.org/10.1088/1755-1315/352/1/012038 Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence (CC BY 3.0)
Paper 3: Zhuravchak, Ruslan; Pedrero, Raquel Alonso; Crespo del Granado, Pedro; Nord, Natasa; Brattebø, Helge. Top-down spatially-explicit probabilistic estimation of building energy performance at a scale. Energy and Buildings 2021 ;Volum 238. https://doi.org/10.1016/j.enbuild.2021.110786 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Paper 4: Zhuravchak, Ruslan; Pedrero, Raquel Alonso; Crespo del Granado, Pedro; Nord, Natasa; Brattebø, Helge. Top-down spatially-explicit probabilistic estimation of building energy performance at a scale. Energy and Buildings 2021 ;Volum 238. https://doi.org/10.1016/j.enbuild.2021.110786 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).